Who doesmultilingualdata work.

Acolad

Paris, France · Growth

European language services group covering translation, localization, and content services, with AI data services offered alongside.

DataForce / TransPerfect

New York, NY · Subsidiary

TransPerfect's AI training data business: multilingual data collection, annotation, and linguistic validation built on TransPerfect's global translation workforce.

e2f

San Jose, CA · Growth

Language data and localization company supplying multilingual training data, linguistic annotation, and evaluation to AI teams.

Keywords Studios

Dublin, Ireland · Growth

Global services provider to the games and entertainment industry, including localization, audio, and player support, with an AI data and testing line.

LILT

San Francisco, CA · Series C

AI-first translation platform combining adaptive machine translation with human linguists in the loop, and a growing line of multilingual data and evaluation services for model builders.

Lingo.dev

Developer-focused localization tooling that uses LLMs to translate apps and content from the codebase.

Lionbridge

Waltham, MA · Growth

Large translation and localization provider that also offers AI training data, evaluation, and model testing services across hundreds of languages.

Pangeanic

Valencia, Spain · Growth

Translation and language technology company offering machine translation, anonymization, and multilingual datasets for AI training.

Phrase

Prague, Czechia · Growth

Localization platform (formerly Memsource) covering translation management, machine translation orchestration, and quality scoring.

Powerling

France

Localization and global content services company (25+ years, 150+ experts, 6 offices, B Corp) that also builds custom speech and audio datasets for AI, including dual-channel conversational recordings and low-resource languages.

RWS TrainAI

Chalfont St Peter, UK · Public

The AI data services arm of RWS, one of the largest language services companies. Offers multilingual data collection, annotation, and linguistic evaluation drawn from its translator and linguist network.

Smartling

New York, NY · Growth

Cloud translation management platform with integrated machine and human translation services.

Translated

Rome, Italy · Growth

Language services company behind ModernMT and a large professional translator network, combining adaptive machine translation with human post-editing.

Unbabel

Lisbon, Portugal / San Francisco, CA · Series C

Language operations platform that pairs machine translation with a community of human editors, plus translation quality estimation used to evaluate MT output.

wxrks

Translation management system with native AI, built to unify cost, quality, workflow, and automation across localization programs.

Appen

Sydney, Australia · Public

One of the longest-running crowd data companies, with roots in linguistic data: collection, annotation, and evaluation across audio, text, image, and video in many languages.

Argos Data

Kraków, Poland · Subsidiary

AI data arm of language services provider Argos Multilingual: multilingual evaluation, annotation, and data collection including lower-resource languages.

Centific

Bellevue, WA · Growth

Large-scale data services with a multilingual workforce; OneForma is its crowd platform for collection, annotation, and evaluation work.

DATAmundi

Multilingual AI data services covering evaluation, annotation, and language data. Limited public information.

Hume AI

New York, NY · Series B

Emotionally intelligent voice AI research lab, known for expressive speech models and large-scale human ratings of vocal and facial expression across cultures.

Innodata

Ridgefield Park, NJ · Public

Data engineering, annotation, and model safety evaluation for large AI labs, with a large multilingual delivery workforce.

Invisible Technologies

New York, NY · Growth

Managed AI training data operations with a large trained workforce; a major post-training (RLHF and SFT) vendor for frontier labs, including multilingual programs.

OpenTrain

Talent network of 476,000+ pre-vetted AI trainers and data labelers in 230+ countries for RLHF, LLM evaluation, red-teaming, and annotation, including multilingual evaluation across 118+ languages.

Perle AI

San Francisco, CA · Seed

Expert-driven data labeling and evaluation with deep expertise in linguistics, medical, and robotics data, built on verified human contributors with provenance tracking.

Prolific

London, UK · Series A

A vetted, demographically balanced participant pool originally built for academic research, now used for human evaluation, preference data, and model testing.

PublicAI

Web3-based train-to-earn network where contributors complete data tasks to improve AI models and earn rewards.

Respondent

Verified research-participant panel (4.3M+ people across 150+ countries) with an AI training line: model evaluation and safety, red-teaming, alignment and preference data, and multilingual, cross-cultural work in 100+ languages.

Scale AI

San Francisco, CA · Growth

Large generalist data engine covering RLHF, SFT, evaluation, and public leaderboards, with multilingual contributor programs through Outlier and Remotasks.

Surge AI

San Francisco, CA

High-skill human data for frontier labs: RLHF, SFT, evaluation, and red-teaming with vetted expert annotators, including multilingual programs.

TaskUs

New Braunfels, TX · Public

Outsourced digital services company with an AI services line covering data labeling, RLHF, safety, and red-teaming across many languages.

Toloka

Amsterdam, Netherlands · Growth

Global crowd and expert network for data collection, annotation, and evaluation, with multilingual benchmark and red-teaming programs for frontier labs.

Welo Data

New York, NY · Subsidiary

Data services arm of localization firm Welocalize, offering annotation, evaluation, and expert networks across 250+ languages.

Besimple AI

Human data and annotation for AI, including speech and audio. Limited public information.

David AI

San Francisco, CA

Audio data company building large conversational speech datasets for voice AI and speech model training.

Defined.ai

Seattle, WA · Series B

Marketplace and services for ethically sourced AI training data (formerly DefinedCrowd), strongest in speech and multilingual datasets.

Fluffle

Consumer app that pays people to talk with their friends, generating natural conversational speech data.

Liva AI

San Francisco, CA · Seed

Building socially intelligent AI, with a focus on human voice and expression data. Founded by Ashley Mo and Aoi Otani.

LXT

Toronto, Canada · Growth

AI data company specializing in speech and multilingual data collection, transcription, and evaluation across 1,000+ language locales.

oto

Full-duplex speech data for voice AI: datasets, models, evals, and tools built from real two-channel conversation.

Panels

Speech and voice data collection for AI. Limited public information.

Sigma AI

Madrid, Spain · Growth

Human data company with long experience in speech and linguistic annotation, transcription, and multilingual data collection.

Silencio AI

Voice and audio data network for AI, sourcing real voices from people in 180+ countries, including languages current voice AI handles poorly.

Sonexis

India

Speech and audio data services. Limited public information.

SpeechData.ai

Speech data collection and transcription for ASR and TTS. Limited public information.

TELUS Digital

Vancouver, Canada · Public

Global data collection and annotation at BPO scale, with a large multilingual crowd for speech, text, and cultural data.

The Agentic Data Co.

San Francisco, CA · Seed

Training data for speech models. Founded by Christian Vestergaard and Matias Voldby Drejer.

Uplift AI

Voice AI for underserved languages, building speech models and data for languages such as Urdu.

Argilla

Madrid, Spain · Subsidiary

Open-source collaboration tool where AI engineers and domain experts build and curate datasets for language-model fine-tuning, RLHF, and evaluation, now part of Hugging Face.

Chipo

Human data for AI training. Limited public information.

Concentrix

Newark, CA · Public

Global customer-experience BPO with AI data services, including annotation and RLHF delivered in many languages.

Lifewood

Hong Kong

AI data services company with delivery centers across Asia and Africa, covering annotation, collection, and multilingual data.

Luel

San Francisco, CA · Seed

Turns everyday words and actions into usable training data for AI (per its YC profile).

Mundo AI

Vancouver, Canada · Seed

High-quality multilingual training data for AI models (per its YC profile).

Turing

Palo Alto, CA · Series E

Post-training data and services for frontier labs, specializing in coding, reasoning, and SFT/RLHF data from a global network of vetted experts.

Uber AI Solutions

San Francisco, CA · Public

Uber's data business: annotation, collection, and human-in-the-loop work drawn from its global driver and courier network plus a dedicated expert workforce across many languages.

UsergyAI

Human feedback and annotation for AI. Limited public information.

Cognegica Networks

Language and cultural data services. Limited public information.

Kalasko

Language and cultural data services. Limited public information.

Karya

Bengaluru, India · Nonprofit

Nonprofit data company that pays rural and low-income workers in India to create speech and text data in Indian languages, with workers sharing in data resale.

Nerval AI

Data for underserved languages and regions. Limited public information.

XRI Global

Language technology and data for low-resource languages. Limited public information.